[ih] RFC 1918 addresses

[Noel Chiappa]

    > From: Craig Partridge

    > As I recall, 10/8 was because it was the only prefix still around of
    > that size (having until recently been the ARPANET's IP network number).

I'm not sure about that - I think there were still A's around at the time
(although being hoarded).

But it was definitely picked because it _had_ been the ARPANet's number. There
was a lot of concern that even though the ARPANet was, by then, gone, various
ARPANet addresses were hard-wired into code, hither, thither and yon, and it
was felt that making net 10 addresses be 'local' would kill two birds with one
stone.

    > I suspect similar reasons drove the other two

Wasn't the C address something Sun had used? Or did it become common through
Sun after it was made the C address?

Probably a good place to look is in the later 'Assigned Numbers' RFCs, and see
if they had any use prior to asssignment to this.

    >> the prefixes defined in RFC 1918 ... 172.12/12, 192.168/16)

?? I had this bit set that they were an A, a B (/16) and a C (/24)? Or maybe
it was a block of B's and C's, resulting in the /12 and /16?

   Noel